NAD+ side effects and safety context

Nearly all the controlled human safety data for the NAD+ category come from oral precursor trials — nicotinamide riboside and nicotinamide mononucleotide — running six to twelve weeks. The intravenous NAD+ infusion, which is the form people are usually asking about when they search for side effects, has almost no randomized safety literature at all. Reading precursor tolerability as though it applies to a clinic drip is the most common error in this area, and it runs in the reassuring direction.

The best-characterized dataset is a 2×6-week randomized, double-blind, placebo-controlled crossover trial of 500 mg nicotinamide riboside twice daily (1,000 mg/day) in healthy middle-aged and older adults, 30 randomized and 24 completing, published in Nature Communications (Martens et al., 2018; PMID 29599478). The authors reported NR well tolerated at that dose with no serious adverse events; the two dropouts attributed to side effects both occurred during the placebo phase.

Two caveats limit how far this carries. The sample was small and healthy, and the duration was six weeks per arm — which says nothing about a year of daily use. And a nicotinamide-based precursor is not plain niacin: high-dose nicotinic acid is separately well known for causing skin flushing, which is a different compound with a different profile.

The effects most commonly described with IV NAD+ — flushing, warmth, chest or abdominal tightness, nausea, cramping, lightheadedness, headache — are reported consistently in clinic and community accounts and are widely said to ease when the infusion is slowed. They are also, notably, not documented in a body of randomized trials, because those trials largely do not exist.

The published IV work is small and pharmacokinetic rather than safety-focused. A pilot study in Frontiers in Aging Neuroscience (PMID 31572171) infused NAD+ intravenously at 3 µmol/min over 6 hours specifically because, as the authors put it, no data were then available on the fate of directly infused NAD+ in a human cohort. It found no change in plasma NAD+ or metabolites until after 2 hours, with increased urinary excretion of NAD+ and methylnicotinamide by 6 hours. That is a study of disposition in a handful of people, not a characterization of adverse events across a population.

So the honest framing is: the infusion sensations are real enough to be described consistently, and their frequency, severity, and predictors have not been measured under controlled conditions.

A distinct risk category has nothing to do with NAD+ itself:

• Compounded and gray-market IV products carry sterility and purity risks, and NAD+ appears on FDA lists of substances raising concerns for compounding.
• Any intravenous or subcutaneous administration adds the ordinary risks of the procedure — local irritation, and infection where sterile technique fails.
• Interaction data are limited across the whole category, so caution alongside prescription medicines is commonly advised rather than demonstrated to be unnecessary.
• People who are pregnant or breastfeeding, or who have significant medical conditions, sit outside the populations these small trials enrolled.

Regulatory status compounds this. NAD+ and NMN are not FDA-approved drugs for any indication, and NMN's dietary-supplement status has been contested in the United States because it was investigated as a drug — which means no labeled product carries a reviewed safety section.

"Well tolerated" in the NR literature means: at 1,000 mg/day, orally, in about two dozen healthy adults, over six weeks, no serious adverse events were seen. It does not extend to higher doses, to injection, to infusion, to years of use, or to people with conditions the trials excluded. The Covarrubias review in Nature Reviews Molecular Cell Biology (NAD+ metabolism and its roles in cellular processes during ageing) states plainly that much remains unknown about whether NAD+ repletion is safe and beneficial in ageing humans — which, coming from the field's own synthesis, is the appropriate calibration.